Meta faces privacy concerns over human-operated call service for AI assistant muse

Meta has faced internal privacy concerns after testing a human-operated call service for its AI assistant Muse. The company, which recently launched the tool, allowed human contractors to handle some phone calls made by Muse, raising alarms among employees about potential data leaks. A Meta vice president acknowledged the testing was a mistake and stated the feature had been temporarily rolled back. Employees raised concerns about sensitive information being shared with third-party contractors, with one noting a racist comment made by a human operator during a call. Muse, which has topped the US app charts, has seen over 2.5 million downloads since its launch. Meta’s stock has risen over 20% since the release, with a market value increase exceeding $200 billion. The company emphasized security measures, including isolated virtual machines for each user session, but internal critics warned these could be undermined by the human intervention.